PERFORMANCE ANALYSIS OF LOGIC COVERAGE CRITERIA USING MUMCUT AND MINIMAL MUMCUT

Abstract

Logic coverage criteria are central aspect of programs and specifications in the testing of software. A Boolean expression with n variables in expression 2 n distinct test cases is required for exhaustive testing .This is expensive even when n is modestly large. The possible solution is to select a small subset of all possible test cases which can effectively detect most of the common faults. Test case prioritization is one of the key techniques for making testing cost effective. In present study performance index of test suite is calculated for two Boolean specification testing techniques MUMCUT and Minimal-MUMCUT. Performance index helps to measure the efficiency and determine when testing can be stopped in case of limited resources. This paper evaluates the testability of generated single faults according to the number of test cases used to detect them. Test cases are generated from logic expressions in irredundant normal form (IDNF) derived from specifications or source code. The efficiency of prioritization techniques has been validated by an empirical study done on bench mark expressions using Performance Index (PI) metric.